search

ProtProtocols / IsoProt

Protocol of the analysis of iTRAQ/TMT proteomics data including quantification, statistical analysis and maybe clustering
https://protprotocols.github.io
11 stars 4 forks source link

Add parameter for peptide inference #45

Open veitveit opened 6 years ago

veitveit commented 6 years ago

Simple options would be: a) grouping of shared proteins (this is happening now) b) only unique peptide

Future options could/will be different versions of the parsimony principle

jgriss commented 6 years ago

@veitveit

Does this apply to how peptides are grouped or how protein inference is done?

veitveit commented 6 years ago

Grouping

On Mon, 6 Aug 2018, 11:08 Johannes Griss, notifications@github.com wrote:

@veitveit https://github.com/veitveit

Does this apply to how peptides are grouped or how protein inference is done?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/ProtProtocols/IsoLabeledProtocol/issues/45#issuecomment-410640901, or mute the thread https://github.com/notifications/unsubscribe-auth/APEZhRIPOZYkEMRAaPn_69qgkZLzV19Kks5uOAd-gaJpZM4VXjN6 .

jgriss commented 6 years ago

@veitveit I've added the GUI element to my local working copy but I'm struggling to implement the actual functionality to only use unique peptides.

Currently, the grouping is done by peptide shaker. The protein groups are automatically generated so in the MSnbase object all our peptides have nprot = 1 (meaning a unique matching).

veitveit commented 6 years ago

@jgriss In MSbase, on can use the grouBby option in the combineFeatures function. We have scripts to do redo the protein grouping by e.g. parsimony principle. So this is in principle doable without too much effort.

When reprocessing data, having the option will help to compare IsoLabeledProtocol results to the ones done in the original studies.

But I will move the milestone to 0.2 as this is not that urgent.

jgriss commented 6 years ago

@veitveit

I think that it will require a bit more code since MSnbase is currently treating a protein group as one protein. But I fully agree that this is doable and will be a very sensible feature. I'll add the GUI element just without any option for now.

jgriss commented 6 years ago

@veitveit The drop-down box is there, but currently only displays a single option. Do you want to keep this issue open until we have the functionality in the R code as well?

veitveit commented 6 years ago

@jgriss yes, let's leave it open until we have a first solution for an alternative grouping